LLM recommendation website traffic has been expanding + 65 % year-to-date. However we need to presume 0 in the future.

LLM Referral Traffic Is Diminishing

LLM recommendation web traffic in B 2 B expanded + 65 1 % considering that January– but dropped – 42 6 % because July

My December forecast of 50 % organic by 2027 is dead:

• In December 2024, I analyzed six B 2 B websites and found LLM recommendation web traffic was expanding at such a rapid rate it would make up 50 % of organic traffic in 3 years.
• Today, I’m locating the month-to-month development price of LLM web traffic went down from 25 1 % in 2024 to 10 4 % in November 2025
• Also from January to July 2025, the average development price was reduced (19 2 %) than my estimate. That’s quickly, however insufficient to reach 50 % natural website traffic in 3 years.

LLM contribution to natural traffic grew from 0. 14 % in 2024 to 1 10 % in 2025, which is more than I projected (0. 79 %).

Yet with organic traffic falling due to AI Overviews , this growth ends up being worthless.

Fewer Citations Despite Expanding Usage

In August, a number of aspects influenced LLM reference web traffic:

1. Seasonality: Siege Media recorded that B 2 B websites lost LLM traffic in August due to holiday period.
2. Router: ChatGPT 5, which introduced on August 7, has a router that chooses the design. The router favors non-reasoning designs, which reveal fewer citations and send much less website traffic out.
3. Focus: Josh from Extensive discovered a higher concentration of references to Reddit and Wikipedia starting late July.

Company seasonality has a reduced impact because neither ChatGPT (customer focus) nor Claude (organization emphasis) sees a decline in website visits

ChatGPT discusses, nonetheless, stopped by one-third in October and continue decreasing in November.

Citations for huge domains like Reddit or Wikipedia do the same (based upon Profound data).

Verdict: LLM sees are up, which removes seasonality as leading reason. The driver of reduced recommendation web traffic is ChatGPT, showing less citations as a result of the design router.

Visibility Is The Genuine Cost

Website traffic was never ever the right way to value LLMs because LLMs make clicks repetitive:

• The AI Mode study I released last month confirms that clicks just taken place for shopping-related jobs (zero-click share = ~ 100 %).
• Church bench Study has found that only 1 % of customers click web links in AI Overviews.

Focusing on traffic causes unsatisfactory results. ChatGPT is a lot more like TikTok than Google Browse. The money of the AI world is presence.

The bright side: LLMs grow the pie. Semrush located individuals don’t make use of Google less frequently because they likewise utilize ChatGPT. If LLMs are additive to Google Browse, the presence surface grows despite the fact that clicks per resource diminish. You have more places to be seen, fewer clicks per location.

However our success metrics need to alter. Referral website traffic neither benefits ChatGPT neither Google, as AI Overviews and AI mode swallow even more clicks. Instead, we need to take on visibility-first.

Default To Absolutely No LLM Traffic

1. Track LLM and natural search seasonality for your upright to gauge the total pie of citations and understand drops/spikes.
2. Display complete citation and state count to address the inquiry, “Are we expanding since the market grows?” Lower citations/mentions implies less chances to influence acquisition decisions.
3. Prioritize brand mentions over citations in LLMs. References without links drive knowledge and influence acquisition choices.
4. Stop expecting (meaningful) LLM reference website traffic. Allocate exposure.
5. Spend resources where LLMs go to train: UGC and third-party reviews like Reddit, YouTube, testimonial websites, community forums.

Featured Image: Paulo Bobita/Search Engine Journal